Chinese AI startup DeepSeek has just officially released its latest large language model (LLM), DeepSeek-V3-0324. With a capacity of up to 641GB, this model was announced on the Hugging Face AI platform without much information being revealed, in line with the company's usual secretive style about new products.
What makes this model unique is that it is licensed under the MIT license, which allows it to be used freely for commercial purposes. Initial benchmark results show that DeepSeek-V3-0324 is capable of running on mainstream hardware configurations, such as Apple’s Mac Studio with an M3 Ultra chip. AI scientist Awni Hannun reported that it was possible to achieve a processing speed of more than 20 tokens per second using this configuration. This ability to run a large language model on commodity hardware on-premises is a stark contrast to the traditional approach of using massive data center infrastructure to support advanced AI models.
According to DeepSeek, initial tests have shown significant improvements over previous versions. The model has been rigorously tested by internal stakeholders and has demonstrated excellent performance, even outperforming all competing models and beating Anthropic’s Claude Sonnet 3.5 on tasks that do not require deep learning. However, unlike Sonnet-style models that require a paid subscription, DeepSeek-V3-0324 is completely free to download and use.
Technically, DeepSeek-V3-0324 uses a mixture of experts (MoE) architecture. It is notable for its selective use of approximately 37 billion of the 685 billion parameters for each task, which increases efficiency by reducing computational demands while maintaining performance. The model also applies Multi-Head Latent Attention (MLA) and Multi-Token Prediction (MTP) technologies, which contribute to improved context memory and faster output.
Users can access DeepSeek-V3-0324 via Hugging Face, OpenRouter's chat interface and API, and DeepSeek's chat platform if they wish. Inference service provider Hyperbolic Labs also provides access to the model.
With the release of DeepSeek-V3-0324, the company continues to assert its position in the race to develop large language models, while providing a powerful and accessible option for the AI research and development community. The ability to run on commodity hardware, coupled with a free-to-use license, will certainly open up many opportunities for practical applications for this model in the near future.
In January, Microsoft announced plans to bring NPU-optimized versions of the DeepSeek-R1 model directly to Copilot+ computers running on Qualcomm Snapdragon X processors.
China-based AI app DeepSeek is currently leading the app store charts.
Bad guys are creating thousands of DeepSeek-like websites in the hopes that unsuspecting users will give them their personal information.
Japanese investment giant Softbank is planning to invest between $15 billion and $25 billion in OpenAI. If the deal goes through, Softbank will become OpenAI’s largest investor, replacing Microsoft.
A series of DeepSeek scam ads on Google are delivering dangerous information-stealing malware to unsuspecting DeepSeek users.
China-based artificial intelligence startup DeepSeek is planning to be more “transparent” about the technology behind its open-source AI models, such as the R1 reasoning model.
Huawei is embarking on a completely Android-free era with the Pura X, a new foldable phone that runs the company's HarmonyOS 5 software and doesn't include an option for people to use Google's operating system.
DeepSeek is the name of a Chinese artificial intelligence company and its eponymous chatbot. The company's flagship models, DeepSeek-V3 and DeepSeek-R1, have made waves because they are cheaper to train and use than comparable models.
The past week has not been a happy one for Nvidia investors or fans.
DeepSeek, the Chinese company behind the free AI model that is shaking up the tech world, is facing a number of negative allegations, including the use of smuggled Nvidia GPUs to train and run its eponymous AI model.
Whether Alibaba's claims will come true remains to be seen, but it looks like ChatGPT and DeepSeek now have a worthy new competitor.
Genshin Impact 5.5 Code helps you exchange for Primordial Stones, Magic Minerals, experience and many other rewards.
macOS Catalina and iPadOS include support for a new feature called Sidecar, designed to let you use your iPad as a secondary display for your Mac.
This is a set of Nokia 1280 wallpapers, if you have ever texted to get 1280 wallpapers, brick phone wallpapers, then try looking at these wallpapers.
Venus Aerospace has revealed the first images of its new hypersonic aircraft called the Stargazer, which can reach a top speed of 11,115 km/h, equivalent to Mach 9.
Bloating can happen to anyone. The good news is that breakfast is a great time to add a few ingredients to your meal that can help reduce bloating. That ingredient is bananas.
OpenAI has just officially introduced a remarkable upgrade to the AI image generation capability in ChatGPT, an important step forward instead of using a separate image generation model like the previous DALL-E.
Chinese AI startup DeepSeek has just officially released its latest large language model (LLM), DeepSeek-V3-0324.
Gboard has long been the go-to keyboard on Android phones. However, many people have recently switched to Samsung Keyboard and haven't looked back.
Send sweet and romantic weekend wishes to your lover. Although they are just words or messages, they will be meaningful spiritual gifts, helping to strengthen your relationship.
HBO's Harry Potter remake is expected to premiere in 2026. Here's what makes the TV Harry Potter a success.
Beautiful moon pictures are not rare on the Internet. This article will summarize for you the most beautiful moon pictures in the world.
Now, AMD has responded to the issues with its underperforming flagship CPU line with a new update that could deliver a few more frames per second.
In October 2022, Elon Musk spent $44 billion to acquire X, but the platform's value has now dropped 78.6% to about $9.4 billion.
When NASA's Artemis I mission successfully launches into space and begins its journey around the moon in 2022, there won't be any astronauts on board — instead, there will be two adult-sized mannequins.
Cleaning doesn't have to be a chore if you know how to make it fun. Here are some simple ways to make cleaning fun.
